Purpose: With the recent discovery that α-methlyacly-coenzyme A racemase (AMACR) is over expressed in a majority of prostate cancer (CaP) specimens we developed a novel polymerase chain reaction (PCR) based approach that would predict the presence of CaP from prostatic secretions. Materials and Methods: A total of 21 patients were enlisted in this study, including 10 with CaP, 2 with high grade PIN and 9 cancer-free individuals (7 healthy men and 2 with benign hyperplasia). Total cellular RNA was extracted from prostatic secretions obtained from post-massage urine specimens. Levels of AMCAR transcripts and prostate specific antigen (PSA) transcripts in these samples were determined by quantitative reverse transcriptase-PCR analyses. Relative AMACR value scores (RAVSs) were calculated by normalizing the AMACR transcript level to that of PSA for each sample and multiplying by 100. An experimentally defined diagnostic cutoff RAVS value was determined in the cancer-free control group. Results: Neither AMACR nor PSA mRNA levels were predictive of CaP when used alone. However, using RAVS values and imposing a diagnostic cutoff value of 2 SDs above the mean RAVS in the cancer-free control group all 9 (100%) cancer-free individuals, including those with benign prostatic hyperplasia, were below the cutoff and 7 of 10 (70%) with CaP had RAVS above the cutoff. Furthermore, 2 of the 3 false-negative cases showed clinically insignificant disease. The 2 patients with high grade PIN were above the cutoff in this test. Conclusions: In this study the quantification of AMACR transcripts normalized to PSA transcripts in prostatic secretions was shown to be predictive of CaP. Therefore, our novel approach using quantitative reverse transcriptase-PCR to detect the AMACR-to-PSA transcript ratio shows promise as a noninvasive screening test for CaP. Furthermore, early results demonstrate a trend toward excluding patients with clinically insignificant disease that may not yet require aggressive treatment due to a low cancer burden.